lectures.alex.balgavy.eu

Lecture notes from university.
git clone git://git.alex.balgavy.eu/lectures.alex.balgavy.eu.git
Log | Files | Refs | Submodules

commit f2680abdfbc3434ee62ba86d97062027a012959b
parent 5a6b02eefb94e16c8d7a7f6e901efe35e4b27075
Author: Alex Balgavy <alex@balgavy.eu>
Date:   Tue,  2 Nov 2021 14:24:05 +0100

Add ACN notes

Diffstat:
Mcontent/_index.md | 1+
Acontent/acn-notes/_index.md | 6++++++
Acontent/acn-notes/intro.md | 30++++++++++++++++++++++++++++++
3 files changed, 37 insertions(+), 0 deletions(-)

diff --git a/content/_index.md b/content/_index.md @@ -9,6 +9,7 @@ title = "Alex's university course notes" * [Advanced Operating Systems](aos-notes/) * [Software Security](softsec-notes/) * [Hardware Security](hwsec-notes/) +* [Advanced Computer Networks](acn-notes/) ## Subject notes: Year 1 * [Computer & Network Security](computer-network-security/) diff --git a/content/acn-notes/_index.md b/content/acn-notes/_index.md @@ -0,0 +1,6 @@ ++++ +title = 'Advanced Computer Networks' ++++ + +# Advanced Computer Networks +1. [Intro](intro) diff --git a/content/acn-notes/intro.md b/content/acn-notes/intro.md @@ -0,0 +1,30 @@ ++++ +title = 'Intro' ++++ + +# Intro + +Circuit switching: +- physical channel with stream of data from source to destination +- three phases: set up, data transfer, tear down +- data transfer involves no routing +- have to reserve circuits + +Packet switching: +- message split into short packets, each handled separately +- one operation - send packet +- packets queued in each router, forwarded to appropriate neighbor + +Bandwidth: max rate of data transfer across given network path + +Latency: amount of time taken to deliver data from source to destination + +Performance metric - flow completion time +- how long does it take to complete traffic flow? +- how long does it take to complete a set of correlated flows? + +Network reliability: +- end-to-end argument: "if a function can only be correctly implemented end-to-end, it must be implemented in the end systems." +- fate-sharing principle: to deal with potential failures, store critical system state at nodes which rely on that state. Only way to lose the state is if the node that relies on it fails, at which point it doesn't matter. + - if you store on network devices, need state to be cleaned up or recreated +- packet vs circuit switching: packet switching is better than circuit switching in terms of *resilience* and *efficiency*. But circuit switching performance is more predictable.